[OSM-talk] A plea for meaning ful changeset comments

[Steve Bennett]

On Fri, Jul 30, 2010 at 9:18 PM, Frederik Ramm <frederik at remote.org> wrote:
> The other group consists of well-meaning mappers who are valuable members of
> our community but who perceive the need to enter a changeset comment as a
> kind of nagging, nannying, and who might be tempted to enter a useless
> comment as a form of protest against that. I'm sure everyone who has to work
> with version control systems of any sort knows the feeling - change one line
> of code and then have to write two lines of commit comment!

And then there's the group that consists of Potlatch users in Live
mode. Potlatch kind of supports changeset comments, but you have very
little control over when the changeset is created and saved, and you
can't (afaik) change a changeset comment after the fact.

For me, very frequently, the changeset just represents a random bunch
of edits I happened to be doing at one time, with not much cohesion.
There are different suburbs all in the same changeset as I flitted
about.

I also question this "value" you talk about. I don't think I've ever
looked at another member's changeset. If the user interfaces made that
a more common occurrence, I'd probably put more effort into changeset
comments, but for me they're not very visible.

(Corollary: when another user tells me specifically that they would
find my changesets easier to navigate if I commented them properly, I
would re-evaluate. But afaik, no one ever looks at my work, so it
seems a bit pointless.)

Steve